Screening, comprehensive assessment, and effective treatments are necessary but not sufficient to adequately treat individuals with depression. They must be paired with effective and efficient systems for delivering the care. Implementation strategies designed to identify, evaluate, and treat patients wherever they present while seamlessly coordinating their care with other members of the health care team is the challenge. Teams of experts who can communicate with each other and coordinate their therapeutic efforts are essential. Furthermore, a population-based orientation is necessary in order to design systems that are scalable and affordable. Incorporating disease registries allows for efficient monitoring and maximizing patient adherence and optimizing treatment outcomes.¹ This requires the use of computerized technologies to enable tracking and monitoring of symptoms, functioning, and satisfaction over time. Computerized treatment also provides greater access by removing the physical barrier to care and thereby enhancing patient engagement in their care.

In this chapter we discuss the provision of care to medically ill patients with depression, models of depression care delivery in medical outpatient settings, and the components of successful, efficient, and cost-effective delivery models. We also address the role of computerization in delivering care to large populations.

THE MEDICAL OUTPATIENT SETTING

In the late 1970s, the first multi-site mental health epidemiological study in the United States identified the primary care system as the “de facto mental health system” for Americans with more prevalent but less severe mental health disorders, including depression.² That trend has continued to the present day.³

Research has underscored the benefit of identifying and treating depression within the primary care setting and a number of promising care models have been developed. Despite this, mental health care in many primary care settings remains suboptimal. The recent Institute of Medicine report, “Improving Quality of Health Care for Mental and Substance Use Conditions”⁴ documented substantial inadequacies in the provision of mental health care, including poor detection, treatment and follow-up care. These findings have documented the need for a wider dissemination of guideline-driven treatment strategies.

According to a recent review, of the 6% to10% of primary care patients who meet diagnostic criteria for MDD, only 50% are accurately diagnosed by primary care providers (PCPs).⁵ Primary care patients who are given antidepressants often receive little education about depression and have infrequent follow-up visits, leading to poor treatment adherence. Poor adherence is common across all classes of prescribed antidepressants: only 25% to 50% of primary care patients continue with antidepressant treatment for the guideline-recommended duration, and 15% never start the medication.⁵

A number of barriers to adherence have been identified, including inadequate monitoring, insufficient patient education about depression and its treatment, and inadequate clinician knowledge and skills.⁶ Stigma, and legal and financial difficulties are also obstacles.

Integrated care models and treatment strategies, such as collaborative care, have been implemented to bridge these gaps and are promising vehicles to improved and affordable care.⁷ Components of these models that have been considered critical are listed in Table 23-1.

Screening is essential for efficiently identifying patients at risk. The ideal tool for screening is sensitive yet specific, self-administered, and easily integrated into the work flow. The Patient Health Questionnaire (PHQ-9)⁸ has been widely used as the screening tool of choice, though it has not yet universally accepted. (See Chapter 3 for a more complete discussion of screening tools.) Screening is important, but a positive screen is insufficient for a clinical diagnosis of depression. Screening must always be followed by a comprehensive medical and psychiatric evaluation for patients who screen positive.

Algorithmic treatment guidelines have been shown to improve outcomes and reduce costs in a variety of diseases, including depression.⁹ Since a high percentage of patients who do receive treatment for depression are not given adequate antidepressant doses for a sufficient duration,¹⁰ guidelines are invaluable in promoting evidence-based care. These guidelines must be integrated into the PCP’s workflow so as to be visible to the clinician and actionable in real time. Guidelines can also be used to prompt the clinician to consider important comorbidities, decide when to hospitalize or refer to a specialist, select and dose treatment, and measure progress.

Patient registries are an increasingly important means of tracking patient populations. They allow clinicians to periodically review the progress of individual patients, detect trends within their patient population, and identify patients who are not improving with treatment. Ideally, registries are integrated into the electronic medical record in order to eliminate redundant data entry. These registries allow clinicians and administrators to evaluate the impact of care on outcomes of both individuals and populations, and then iteratively modify programming to improve the quality and efficiency of care. If linked to claims databases, registries make it possible to determine the effects of care on health care costs.

Mental health consultation is the most common and long-standing dimension of integrated care. The consultant is most often a psychiatrist, but may be a nurse, psychologist, or social worker with mental health expertise. Consultants communicate with the PCP in person, by phone, or via email or clinical messaging. Depending on the patient’s complexity and the comfort level and experience of the consultee, the questions may be answered informally through a “curbside” consultation. In other instances, a formal interview of the patient is necessary. Pharmacotherapy is generally provided by the PCP with additional consultation as needed. If the patient or pharmacotherapy is too complex, the patient might be treated by the consultant or otherwise referred to a psychiatric outpatient setting. This approach allows more patients to receive optimal care for their depression in the primary care setting. This reduces the need to refer the patient to a psychiatric outpatient setting, in which case at least half of patients never complete the referral and there is a high rate of drop out among those who do.¹¹ Patient factors (such as forgetting the appointment, interference by the psychiatric symptoms), complicating psychosocial issues (such as homelessness, poor insight, or negative view of psychiatry), as well as clinician factors (such as poor communication, lengthy wait for appointment, and perception of referring clinician devaluing psychiatry) all contribute to the problem of attrition and insufficient care.¹²

The Depression Care Manager (DCM) is often a nurse or social worker with mental health expertise, though a trained bachelors-level person can also be a highly effective DCM. This individual is key in the collaborative care model as he or she assists the PCP in educating the patient about depression and its treatment, reinforcing basic but important activities such as healthy diet and exercise, and tracking and measuring the progress of the patient. The DCM is often trained in motivational interviewing techniques and delivering behavioral activation or problem solving therapy to the patient. Finally, the DCM assists with referrals to therapists when psychotherapy is indicated.

The co-location of mental health clinicians in the primary care setting is another element of integrated care. This provides more opportunities for real-time consultation, increases the opportunities for educating PCPs, and improves the likelihood of a successful mental health referral, since patients receive their specialized care in the familiar primary care setting. The embedded psychiatrist is not only co-located, but is also integrated into the work flow of the primary care team in a standardized, often protocol-driven fashion.

In a stepped care approach, the severity and complexity of the patient’s depression determines the level of care. For example, an individual diagnosed with a mild depression might be provided a list of self-help options available in books or over the Internet. An individual with a more severe, acute, or complicated depression (e.g., one associated with other major psychiatric disorders), might be assigned to a psychiatrist for management. Those in between might be co-managed by both the PCP and DCM. Depending on progress, a patient can be reassigned within this spectrum of care. This model allows for limited resources to be distributed or triaged in the most efficient manner.

Collaborative care is an amalgam of all of the components above and truly embraces the team concept for delivering behavioral health care in the primary care setting. The team is composed of clinicians with behavioral health expertise and is called upon by the PCP when needed. A psychiatrist usually serves as a consultant to the behavioral health team who assesses, educates, suggests additional resources, and often provides skills-based therapies and ongoing monitoring. If a medication is recommended, it is prescribed by the PCP.

A number of clinics and health care systems around the United States have developed or are developing models for integrating mental health care into the primary care setting, incorporating some combination of the elements described above. Here we present two successful and well-established models. We also describe the evolving model at our hospital, which is a hybrid of these two models.

Impact

The Improving Mood–Promoting Access to Collaborative Treatment program, or IMPACT, was the first major collaborative care model to undergo rigorous study and wide dissemination. It is both clinically superior to standard treatment^13,14 and cost-effective.¹⁵ This manualized model relies on a clinical algorithm that is carried out under the direction of a DCM, usually a nurse, social worker or psychologist, in collaboration with the primary care doctor, the consulting psychiatrist, and the patient.

In the IMPACT program, older adult patients identified as depressed or dysthymic by referral from their PCP or by a screening questionnaire, are provided educational materials about depression and a referral to the DCM. The DCM assesses the patient, educates, and serves as a coach, engaging the patient in behavioral activation. The DCM helps the patient navigate the clinical algorithm based on his or her assessment of the patient’s needs and preferences. For most patients, the initial treatment choice is either an antidepressant medication or a short course of problem solving treatment (PST) administered by the DCM. The treatment is then adjusted as needed according to the algorithm and the patient’s progress, monitored by the DCM. For patients who recover, the treatment then focuses on relapse prevention. The consulting psychiatrist reviews cases in which the patient is not responding to treatment and consults on those who pose diagnostic challenges or have a complicated course.¹³

The IMPACT model lends itself well to adaptation in a range of clinical settings.¹³ Comprehensive instructions and tools for implementation of the program are available at the IMPACT website.¹⁶ Thus far, IMPACT’s benefits have been demonstrated largely in older adult patients with depression; efforts to study it in other populations are in early stages.^17,18

Mental Health Integration (Intermountain Health Care)

Intermountain Health Care, a network of hospitals and clinics with a tradition of collaboration between medical teams, has developed another model for depression treatment in primary care, called mental health integration (MHI). MHI does not involve a clinical algorithm but rather stratifies patients based on an initial evaluation and screening (using the PHQ-9). Patients with depression are classified as mild, moderate, or severe and/or complicated, and are then triaged to a corresponding level of care. This may include support staff, care management with or without consultation by a mental health specialist, or referral directly to the mental health specialist, respectively.¹⁹

MHI differs from other models in its emphasis on recruiting family and community support. So far, MHI has been demonstrated to improve the detection of depression without increasing overall health care costs for its patients, and it has been associated with higher levels of patient and provider satisfaction.^20,21

Brigham and Women’s Hospital

The Patient-Centered Medical Home

Brigham and Women’s Hospital (BWH) began developing a collaborative care model in the broader context of creating a Patient-Centered Medical Home (PCMH). Before describing the BWH model, it is necessary to understand the PCMH and its origins. PCMH aims to coordinate health care in a patient-centered, technology-enabled manner. The Joint Principles of the Patient-Centered Medical Home outlines seven key principles that appear in Box 23-1.²²

The National Committee for Quality Assurance (NCQA) developed these general principles into specific measurable standards for designation as a PCMH.²³ Their standards and guidelines are divided into six main areas:

• 
  

  Access and Continuity—including after hours and electronic access, and provision of culturally and linguistically appropriate services

  
  
• 
  

  Identify and Manage Patient Populations—including using registries to proactively remind patients of overdue care

  
  
• 
  

  Plan and Manage Care—including implementing evidence-based guidelines using point-of-care reminders, identifying high-risk patients, and managing medications

  
  
• 
  

  Provide Self-Care Support—including providing educational resources, referring to community resources, providing self-management tools, and creating self-management plans with the patients and their families

  
  
• 
  

  Track and Coordinate Care—including testing and referral tracking and managing care transitions

  
  
• 
  

  Measure and Improve Performance—including patient experience of care

The PCMH model was originally designed to be supported by a blended payment model, in which fee-for-service is supplemented by a capitated fee (per-patient-per-month) to cover additional services that are not covered by current fee-for-service plans, such as care management, preventive interventions, and between-visit communication. PCMH is also considered to be the foundation for the successful implementation of Accountable Care Organizations (ACO).²⁴

Successful management of depression is imperative once a system adopts a bundled care model and becomes financially responsible for total cost of care, as in an ACO. For example, many studies have shown that control of diabetes is poorer and care more costly in patients with comorbid depression compared to nondepressed diabetics.²⁵ Adoption of the Intermountain Mental Health Integration model produced savings of $667 per patient per year in medical expenses, including a 54% decrease in emergency department use.²⁰

Although there is no explicit mention of behavioral health within the Joint Principles, such integration is congruent with its principles of comprehensiveness and coordination of care. The 2011 NCQA standards and guidelines require a practice to implement evidence-based guidelines for patients with three important conditions, including one related to unhealthy behaviors or mental health or substance abuse.

The South Huntington Model

It was clear from the outset that the BWH’s PCMH model had to include behavioral health services. Because the population we serve has a high rate of trauma-related disorders, bipolar, and psychotic disorders, we incorporated a comprehensive psychiatric assessment early in the process. We wanted to assure that individuals who eventually receive antidepressant medication were at low risk for bipolar disorder, that treatment took comorbidities such as anxiety disorders and substance use disorders into account, and that patients with complicated psychiatric histories were identified at the outset. Ultimately we created a model that is best described as a hybrid of IMPACT and Intermountain (Fig. 23-1).

The Behavioral Health Questionnaire is an important component in our program. It allows us to collect important diagnostic information and to screen for other common comorbid conditions. The components of the questionnaire are reviewed in Box 23-2.

The process of gathering this information fosters the patient’s active involvement in his/her care, and also improves the efficiency of behavioral health team meetings at which we decide whether the patient should be managed in the primary care setting and if so, to develop the treatment plan. The questionnaire also serves as a training tool for our clinicians. This enables them to better implement treatment recommendations. Furthermore, it reduces the need for face-to-face consultations with the psychiatrist, thereby allowing him/her to focus on the most challenging cases (Fig. 23-1).

It can be challenging for some PCPs to shift from the traditional dyadic approach to patient care to team-based care. Fundamental to making this transition is clear role delineation, seamless communication, and trust in the other team members. PCPs learn to slow down and not feel compelled to issue antidepressant prescriptions upon first diagnosing a depression, much as antihypertensives are not routinely prescribed after a single high blood pressure reading. Although the team-based care is not embraced by all patients, most patients report very positive experiences. The feedback about the Behavioral Health Questionnaire from patients has been largely positive. They expressed relief that the team wants to be thorough and is actually interested in the nuances of their mental health.

The Emergency Department

Prevalence of Depression in the Emergency Department

In the United States, more than six million patients with mental health conditions are seen each year in emergency departments (EDs).²⁶ There has been a 15% increase in psychiatric diagnoses made in EDs between 1992 and 2000,²⁷ as individuals with psychiatric conditions and no source of primary care increasingly present to EDs.²⁶ The prevalence of depression among patients in EDs is very high. Patients with depression are almost three times as likely to use emergency department services as those without depression, after adjusting for age, race, gender, and comorbid medical conditions.²⁸ In a prospective observational study, the prevalence of depression meeting DSM-IV criteria was 22%.^29,30 Another multicenter study estimated the prevalence of depression in EDs at 30%, and it was associated with female gender and lower socioeconomic status.³¹

The rate of suicide is at least eight times higher in people with depression than in the general population.³² Individuals with mental health problems are more likely to seek care outside the mental health system, from clinicians other than mental health professionals. Seventy-five percent of those completing suicide have had contact with medical care providers within the year of their death, often in emergency departments.³³ Thus, it is imperative that ED personnel become more alert to the possibility of suicide in their patients.

As detailed elsewhere in this book, depression is closely linked to medical illness. This relationship is especially salient in the emergency setting. Thus, for example, a study of inner-city ED patients with diabetes revealed that 60% had some depressive symptoms and 20% screened positive for clinical depression.³⁴ Many studies have shown that patients with depressive symptoms or diagnosed depression report more health problems, more pain, and more disability.^35,36,37 Not surprisingly, these patients present to EDs with somatic complaints related to their comorbid medical illnesses, but these complaints are in reality often due to depression.³⁸ It has been widely argued that somatization, or the expression of psychological distress via physical symptoms, is the single most common reason that psychiatric illness goes undetected in medical settings.³⁹ Some studies suggest that depressed, medically ill patients wait longer to be seen⁴⁰ and are assigned a lower-priority triage score in EDs, due to stigma of mental illness.⁴¹